农机化研究2025，Vol.47Issue(11)：55-64,10.DOI:10.13427/j.issn.1003-188X.2025.11.007
边缘计算下采棉头状态监测及异常检测系统设计
Design of Condition Monitoring and Anomaly Detection System of Cotton Picking Head Based on Edge Computing
摘要
Abstract
Aiming at the problems of monitoring parameters,data processing efficiency,real-time performance and accu-rate identification of abnormal data in agricultural machinery monitoring,an edge computing monitoring and anomaly de-tection system for cotton picking heads was designed.Firstly,the overall design scheme of the system was proposed,which took edge preprocessing layer as the core,received multi-source sensor signals in data acquisition layer to realize condition monitoring,deployed anomaly detection algorithm to improve monitoring data quality,and provided good data support for health assessment,fault diagnosis and life prediction model training of cloud application layer.Then,with STM32 microprocessor as the core,the hardware design of power supply circuit,AD7606 digital-analog chip and wire-less communication chip in data acquisition terminal was completed,and the software design of vibration sensor and other signal acquisition was completed.An isolated forest anomaly detection method based on similarity measurement was pro-posed,which included the construction of multi-domain feature matrix,the estimation of anomaly score,and the recogni-tion of anomaly data.In order to verify the validity and reliability of data acquisition and anomaly detection,field experi-ments and simulation signal tests were carried out,and the results showed that:The data acquisition terminal can work stably and meet the real-time and quantitative requirements of condition monitoring data.The proposed anomaly detection method significantly improved the accuracy rate and reduced false positives while maintaining a high recall rate,providing more reliable technical support for intelligent monitoring of agricultural machinery.关键词
采棉头/状态监测/异常检测/边缘计算/嵌入式物联网Key words
cotton picking head/condition monitoring/anomaly detection/edge computing/embedded internet of things分类
农业科技引用本文复制引用
刘书伟,姜宏,章翔峰..边缘计算下采棉头状态监测及异常检测系统设计[J].农机化研究,2025,47(11):55-64,10.基金项目
国家自然科学基金项目(52265016) （52265016）
新疆维吾尔自治区重大科技专项(2022A02010-3) （2022A02010-3）
新疆农机研发制造推广应用一体化项目(YTHSD2022-07) （YTHSD2022-07）
